Abd El Rahman Shabayek

Abd El Rahman Shabayek

Université du Luxembourg

H-index: 13

Europe-Luxembourg

About Abd El Rahman Shabayek

Abd El Rahman Shabayek, With an exceptional h-index of 13 and a recent h-index of 13 (since 2020), a distinguished researcher at Université du Luxembourg, specializes in the field of Computer Vision, Robotics, Omnidirectional vision, Polarization vision, Non-conventional omnidirectional sensors.

His recent articles reflect a diverse array of research interests and contributions to the field:

Multi-Objective Hardware Aware Neural Architecture Search using Hardware Cost Diversity

Hardware Aware Evolutionary Neural Architecture Search using Representation Similarity Metric

Impact of Disentanglement on Pruning Neural Networks

Compression of Deep Neural Networks for Space Autonomous Systems

Establishing a Multi-Functional Space Operations Emulation Facility: Insights from the Zero-G Lab

Face-GCN: A graph convolutional network for 3D dynamic face recognition

Dense and sparse 3D deformation signatures for 3D dynamic face recognition

Deep network compression with teacher latent subspace learning and lasso

Abd El Rahman Shabayek Information

University

Université du Luxembourg

Position

SnT

Citations(all)

690

Citations(since 2020)

556

Cited By

380

hIndex(all)

13

hIndex(since 2020)

13

i10Index(all)

17

i10Index(since 2020)

14

Email

University Profile Page

Université du Luxembourg

Abd El Rahman Shabayek Skills & Research Interests

Computer Vision

Robotics

Omnidirectional vision

Polarization vision

Non-conventional omnidirectional sensors

Top articles of Abd El Rahman Shabayek

Multi-Objective Hardware Aware Neural Architecture Search using Hardware Cost Diversity

Authors

Nilotpal Sinha,Peyman Rostami,Abd El Rahman Shabayek,Anis Kacem,Djamila Aouada

Journal

arXiv preprint arXiv:2404.12403

Published Date

2024/4/15

Hardware-aware Neural Architecture Search approaches (HW-NAS) automate the design of deep learning architectures, tailored specifically to a given target hardware platform. Yet, these techniques demand substantial computational resources, primarily due to the expensive process of assessing the performance of identified architectures. To alleviate this problem, a recent direction in the literature has employed representation similarity metric for efficiently evaluating architecture performance. Nonetheless, since it is inherently a single objective method, it requires multiple runs to identify the optimal architecture set satisfying the diverse hardware cost constraints, thereby increasing the search cost. Furthermore, simply converting the single objective into a multi-objective approach results in an under-explored architectural search space. In this study, we propose a Multi-Objective method to address the HW-NAS problem, called MO-HDNAS, to identify the trade-off set of architectures in a single run with low computational cost. This is achieved by optimizing three objectives: maximizing the representation similarity metric, minimizing hardware cost, and maximizing the hardware cost diversity. The third objective, i.e. hardware cost diversity, is used to facilitate a better exploration of the architecture search space. Experimental results demonstrate the effectiveness of our proposed method in efficiently addressing the HW-NAS problem across six edge devices for the image classification task.

Hardware Aware Evolutionary Neural Architecture Search using Representation Similarity Metric

Authors

Nilotpal Sinha,Abd El Rahman Shabayek,Anis Kacem,Peyman Rostami,Carl Shneider,Djamila Aouada

Published Date

2024

Hardware-aware Neural Architecture Search (HW-NAS) is a technique used to automatically design the architecture of a neural network for a specific task and target hardware. However, evaluating the performance of candidate architectures is a key challenge in HW-NAS, as it requires significant computational resources. To address this challenge, we propose an efficient hardware-aware evolution-based NAS approach called HW-EvRSNAS. Our approach re-frames the neural architecture search problem as finding an architecture with performance similar to that of a reference model for a target hardware, while adhering to a cost constraint for that hardware. This is achieved through a representation similarity metric known as Representation Mutual Information (RMI) employed as a proxy performance evaluator. It measures the mutual information between the hidden layer representations of a reference model and those of sampled architectures using a single training batch. We also use a penalty term that penalizes the search process in proportion to how far an architecture's hardware cost is from the desired hardware cost threshold. This resulted in a significantly reduced search time compared to the literature that reached up to 8000x speedups resulting in lower CO2 emissions. The proposed approach is evaluated on two different search spaces while using lower computational resources. Furthermore, our approach is thoroughly examined on six different edge devices under various hardware cost constraints.

Impact of Disentanglement on Pruning Neural Networks

Authors

Carl Shneider,Peyman Rostami,Anis Kacem,Nilotpal Sinha,Abd El Rahman Shabayek,Djamila Aouada

Journal

arXiv preprint arXiv:2307.09994

Published Date

2023/7/19

Deploying deep learning neural networks on edge devices, to accomplish task specific objectives in the real-world, requires a reduction in their memory footprint, power consumption, and latency. This can be realized via efficient model compression. Disentangled latent representations produced by variational autoencoder (VAE) networks are a promising approach for achieving model compression because they mainly retain task-specific information, discarding useless information for the task at hand. We make use of the Beta-VAE framework combined with a standard criterion for pruning to investigate the impact of forcing the network to learn disentangled representations on the pruning process for the task of classification. In particular, we perform experiments on MNIST and CIFAR10 datasets, examine disentanglement challenges, and propose a path forward for future works.

Compression of Deep Neural Networks for Space Autonomous Systems

Authors

Carl Shneider,Nilotpal Sinha,Michele Lynn Jamrozik,Marcella Astrid,Peyman Rostami Abendansari,Anis Kacem,Abd El Rahman Shabayek,Djamila Aouada

Published Date

2023/4/19

Efficient compression techniques are required to deploy deep neural networks (DNNs) on edge devices for space resource utilization tasks. Two approaches are investigated.

Establishing a Multi-Functional Space Operations Emulation Facility: Insights from the Zero-G Lab

Authors

Miguel Olivares-Mendez,Baris Yalcin,Mohatashem Reyaz Makhdoomi,Vivek Muralidharan,Zhanna Bokal,Miguel Ortiz del Castillo,Vincent Gaudilliere,Leo Pauly,Olivia Borgue,Mohammadamin Alandihallaj,Jan Thoemel,Ernest Skrzypczyk,Arunkumar Rathinam,Kuldeep Rambhai Barad,Abd El Rahman Shabayek,Andreas Hein,Djamila Aouada,Carol Martinez

Published Date

2023

This methods paper outlines the development of the Zero-G Laboratory at the University of Luxembourg, a crucial resource for advancing research in space operations. The primary objective of this laboratory is to meticulously simulate the micro-gravity conditions encountered in space, allowing for comprehensive testing of space-related hardware and software before their deployment in the demanding environment of outer space. The key methods employed in establishing this facility include replicating space-representative infrastructure elements such as realistic lighting conditions, epoxy flooring, and robotic systems mounted on rails. The laboratory integrates its hardware and software over a centralized Robot Operating System (ROS) network. Researchers can conduct hybrid emulations, combining robotic systems with pre-modeled software components to simulate intricate orbital scenarios effectively. Furthermore, this paper serves as a practical guide for laboratory construction. The aim of this project is to assist the research community in establishing similar facilities and fostering advancements in space-related research and technology development.

Face-GCN: A graph convolutional network for 3D dynamic face recognition

Authors

Konstantinos Papadopoulos,Anis Kacem,Djamila Aouada

Published Date

2022/5/26

Face recognition has significantly advanced over the past years. However, most of the proposed approaches rely on static RGB frames and on neutral facial expressions. This has two disadvantages. First, important facial shape cues are ignored. Second, facial deformations due to expressions can have an impact in the performance of such a method. In this paper, we propose a novel framework for dynamic 3D face recognition based on facial keypoints. Each dynamic sequence of facial expressions is represented as a spatio-temporal graph, which is constructed using 3D facial landmarks. Each graph node contains local shape and texture features that are extracted from its neighborhood. For the classification of face videos, a Spatio-temporal Graph Convolutional Network (ST-GCN) is used. Finally, we evaluate our approach on a challenging dynamic 3D facial expression dataset.

Dense and sparse 3D deformation signatures for 3D dynamic face recognition

Authors

Djamila Aouada

Journal

IEEE Access

Published Date

2021/3/8

This work analyses dense and sparse 3D Deformation Signatures to represent 3D temporal deformation instances. The signatures are employed in dynamic 3D face recognition, however, they are applicable in other domains. This is demonstrated for dynamic expression recognition. The pushing need for non-intrusive bio-metric measurements made face and its expressions recognition dominant players in domains like entertainment, surveillance and security. The proposed signature can be computed from 2D, 3D or hybrid input by means of robust 3D fitting. It is computed given a non-linear 6D space representation which guarantees by construction physically plausible 3D deformations. A unique deformation indicator is computed per triangle in a triangulated mesh as a ratio derived from scale and in-plane deformation in the canonical space. These indicators are concatenated densely or sparsely to form the …

Deep network compression with teacher latent subspace learning and lasso

Authors

Oyebade K Oyedotun,Abd El Rahman Shabayek,Djamila Aouada,Björn Ottersten

Journal

Applied Intelligence

Published Date

2021/2

Deep neural networks have been shown to excel in understanding multimedia by using latent representations to learn complex and useful abstractions. However, they remain unpractical for embedded devices due to memory constraints, high latency, and considerable power consumption at runtime. In this paper, we propose the compression of deep models based on learning lower dimensional subspaces from their latent representations while maintaining a minimal loss of performance. We leverage on the premise that deep convolutional neural networks extract many redundant features to learn new subspaces for feature representation. We construct a compressed model by reconstruction from representations captured by an already trained large model. As compared to state-of-the-art, the proposed approach does not rely on labeled data. Moreover, it allows the use of sparsity inducing LASSO …

Salient Object Detection Using Spatially Weighted Multiple Contrast Cues

Authors

Norhan M Saleh,Mohamed Tahoun,Abd El Rahman Shabayek,M-H Mousa

Published Date

2021

Detecting objects that capture visual attention has played a key role in computer vision. In this paper, we present a model that incorporates multiple bottom-up cues depending on the following concepts: connectivity to the outstanding parts where each superpixel will take a saliency score based on its connectivity strength to the enclosed interest regions, global contrast by measuring how every superpixel differs from all superpixels in the image and utilizing regional frequency tuning and center-bias. The final saliency map is produced by integrating the resulted foreground maps of each cue and refining the result with the optimization framework. An extensive experimental evaluation is done on three challenging datasets to evaluate the proposed model using common classification criteria. Our model has the superiority in performance over the other models qualitatively and quantitatively.

3D sparse deformation signature for dynamic face recognition

Authors

Djamila Aouada,Kseniya Cherenkova,Gleb Gusev

Published Date

2020/10/25

This paper proposes a novel compact and memory efficient Sparse 3D Deformation Signature (S3DS) to represent a sparse 3D deformation signal for 3D Dynamic Face Recognition. S3DS is based on a non-linear 6D-space representation that secures physically plausible 3D deformations. A unique deformation indicator is computed per triangle in a triangulated mesh, thanks to a recent 3D Deformation Signature (3DS) that is based on Lie Bodies. The proposed S3DS sparsely concatenates unique triangular indicators to construct the facial signature for each temporal instance. The novel descriptor shall benefit domains like surveillance and security in providing non-intrusive bio-metric measurements. By construction, S3DS is resistant to common security attacks like presentation, template and adversarial attacks. Two dynamic datasets (BU4DFE and COMA) were examined in various sparse concatenation settings. Using …

Improved highway network block for training very deep neural networks

Authors

Oyebade K Oyedotun,Djamila Aouada,Björn Ottersten

Journal

IEEE Access

Published Date

2020/9/24

Very deep networks are successful in various tasks with reported results surpassing human performance. However, training such very deep networks is not trivial. Typically, the problems of learning the identity function and feature reuse can work together to plague optimization of very deep networks. In this paper, we propose a highway network with gate constraints that addresses the aforementioned problems, and thus alleviates the difficulty of training. Namely, we propose two variants of highway network, HWGC and HWCC , employing feature summation and concatenation respectively. The proposed highway networks, besides being more computationally efficient, are shown to have more interesting learning characteristics such as natural learning of hierarchical and robust representations due to a more effective usage of model depth, fewer gates for successful learning, better generalization capacity and …

3D deformation signature for dynamic face recognition

Authors

Djamila Aouada,Kseniya Cherenkova,Gleb Gusev,Björn Ottersten

Published Date

2020/5/4

This work proposes a novel 3D Deformation Signature (3DS) to represent a 3D deformation signal for 3D Dynamic Face Recognition. 3DS is computed given a non-linear 6D-space representation which guarantees physically plausible 3D deformations. A unique deformation indicator is computed per triangle in a triangulated mesh as a ratio derived from scale and in-plane deformation in the canonical space. These indicators, concatenated, construct the 3DS for each temporal instance. There is a pressing need of non-intrusive bio-metric measurements in domains like surveillance and security. By construction, 3DS is a non-intrusive facial measurement that is resistant to common security attacks like presentation, template and adversarial attacks. Two dynamic datasets (BU4DFE and COMA) were examined, in a standard classification framework, to evaluate 3DS. A first rank recognition accuracy of 99.9%, that …

Towards automatic CAD modeling from 3D scan sketch based representation

Authors

Abd El Rahman Shabayek,Djamila Aouada,Kseniya Cherenkova,Gleb Gusev

Published Date

2020

This paper proposes a novel approach to convert a 3D scan to its CAD counterpart. The objective is to extract intermediate sketch planes that well represent the input scan and are close enough to the original design intent. These sketches can then be easily converted into CAD models automatically due to their faithful representation of the input geometry. One objective is to avoid incorporating user/company dependent content in the CAD reconstruction process. The intermediate representation shall be directly supported in any CAD environment to boost the designer’s work without the need of supplementary (model conversion, automatic feature recognition) steps. Nowadays, it is common to digitize an object and reconstruct its geometric primitives. However, this reconstruction contains only geometry. In literature, the final goal might be met by recovering the modeling tree itself, by means of automatic feature recognition, and converting to the proper format of a specific CAD software package. However, the constructed tree and its conversion introduce issues in the reconstruction process. The definition of an exact modeling tree, and the production of a meaningful final CAD model are rather hard to obtain. This imposes a rather inefficient working method, thereby heavily impacting the designer’s modeling skills.

Home-based rehabilitation system for stroke survivors: a clinical evaluation

Authors

Enjie Ghorbel,Renato Baptista,Abdelrahman Shabayek,Djamila Aouada,Maialen Gorostiza Oramaeche,Janire Orcajo Lago,Leire Ortiz Fernandez

Journal

Journal of medical systems

Published Date

2020/12

Recently, a home-based rehabilitation system for stroke survivors (Baptista et al. Comput. Meth. Prog. Biomed. 176:111–120 2019), composed of two linked applications (one for the therapist and another one for the patient), has been introduced. The proposed system has been previously tested on healthy subjects. However, for a fair evaluation, it is necessary to carry out a clinical study considering stroke survivors. This work aims at evaluating the home-based rehabilitation system on 10 chronic post-stroke spastic patients. For this purpose, each patient carries out two exercises implying the motion of the spastic upper limb using the home-based rehabilitation system. The impact of the color-based 3D skeletal feedback, guiding the patients during the training, is studied. The Time Variable Replacement (TVR)-based average distance, as well as the average postural angle used in Baptista et al. (Comput. Meth. Prog …

See List of Professors in Abd El Rahman Shabayek University(Université du Luxembourg)

Abd El Rahman Shabayek FAQs

What is Abd El Rahman Shabayek's h-index at Université du Luxembourg?

The h-index of Abd El Rahman Shabayek has been 13 since 2020 and 13 in total.

What are Abd El Rahman Shabayek's top articles?

The articles with the titles of

Multi-Objective Hardware Aware Neural Architecture Search using Hardware Cost Diversity

Hardware Aware Evolutionary Neural Architecture Search using Representation Similarity Metric

Impact of Disentanglement on Pruning Neural Networks

Compression of Deep Neural Networks for Space Autonomous Systems

Establishing a Multi-Functional Space Operations Emulation Facility: Insights from the Zero-G Lab

Face-GCN: A graph convolutional network for 3D dynamic face recognition

Dense and sparse 3D deformation signatures for 3D dynamic face recognition

Deep network compression with teacher latent subspace learning and lasso

...

are the top articles of Abd El Rahman Shabayek at Université du Luxembourg.

What are Abd El Rahman Shabayek's research interests?

The research interests of Abd El Rahman Shabayek are: Computer Vision, Robotics, Omnidirectional vision, Polarization vision, Non-conventional omnidirectional sensors

What is Abd El Rahman Shabayek's total number of citations?

Abd El Rahman Shabayek has 690 citations in total.

What are the co-authors of Abd El Rahman Shabayek?

The co-authors of Abd El Rahman Shabayek are Björn Ottersten, Aboul Ella Hassanien Ali, Ibrahim Sadek, Djamila Aouada, Oyebade K. Oyedotun, Renato Baptista.

    Co-Authors

    H-index: 100
    Björn Ottersten

    Björn Ottersten

    Université du Luxembourg

    H-index: 87
    Aboul Ella Hassanien Ali

    Aboul Ella Hassanien Ali

    Cairo University

    H-index: 26
    Ibrahim Sadek

    Ibrahim Sadek

    American University of Sharjah

    H-index: 24
    Djamila Aouada

    Djamila Aouada

    Université du Luxembourg

    H-index: 15
    Oyebade K. Oyedotun

    Oyebade K. Oyedotun

    Université du Luxembourg

    H-index: 9
    Renato Baptista

    Renato Baptista

    Université du Luxembourg

    academic-engine

    Useful Links